
The Scale Up Show · 2025-05-26 · 13 min
Key moments - from our scoring
Substance score
25 / 100
Five dimensions, 20 points each
Ryan Staley, an AI transformation specialist for go-to-market teams, puts Genspark - the new super agent generating significant industry buzz - through three real-world tests to determine whether it can genuinely replace existing AI tools. He evaluates Genspark against competitors in three critical use cases: presentation generation (compared against Gamma), research and data extraction (tested against Google's Deep Research and Gemini), and strategic ideation for AI use cases. While Genspark shows promise in ideation, generating content strategies with social media calendars and engagement frameworks, Staley finds it underperforms in execution. The presentation module produces visually creative but inconsistently rendered slides with alignment issues, while the research function burned through 10,000 credits to extract only 10 portfolio company records from Vista Equity before entering infinite loops on missing data fields. Staley concludes that Genspark isn't yet a unified replacement - Gamma delivers superior presentation visuals, Deep Research outperforms on data compilation, and general-purpose LLMs remain better for true ideation. He shares three custom prompts (the Steve Jobs presentation framework, hyper-specific research queries, and super agent use case discovery) that maximize any agent's output.
Genspark created visually creative slides with animations in 5-6 minutes but had inconsistent quality - some slides were strong while others had off-center visuals and missing elements. Gamma produced fewer animations but higher visual quality and stayed truer to the message, winning the comparison for Staley's use case.
Genspark successfully pulled only 10 companies before getting stuck on the 'number of salespeople' field, entering an infinite loop that forced Staley to stop it. It also burned through 10,000 of his 20,000 monthly credits on this single task, compared to Gemini's Deep Research which retrieved the entire portfolio with LinkedIn URLs and acquisition dates.
Genspark performed strongest on ideation - when asked how a CIO transformation company should use a super agent, it generated structured content strategies including content pillars, posting schedules, carousel post ideas, and engagement measurement frameworks organized week-by-week.
While promising, Genspark lacks reliable execution across specialized domains - presentations export poorly to PowerPoint, research gets stuck on missing data with no failsafes, and it loops endlessly rather than gracefully handling incomplete information. Staley recommends keeping specialized tools like Gamma, Deep Research, and general LLMs instead.
Ask the AI tool directly: 'How do I use you best? What will I get the best results with?' This prompts the tool to think through and explain its own optimal use cases, a technique Staley notes doesn't work with traditional tools like Excel or PowerPoint.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode is primarily a screen-share walkthrough that translates poorly to audio; the few substantive observations (Genspark loops on data extraction, credit burn rate) are buried in descriptive narration and filler. Actionable insight is sparse and the conclusion - use specialised tools instead of a general super-agent - is not novel.
I said basically twice. It's identifying like, like the revenue loss annually based on just the time burn of not implementing and integrating AI in your go to market team
you can't ask Excel how to use Excel better. You can't do the same thing with PowerPoint either
The content is a standard AI-tool comparison with no contrarian or first-principles framing; the Steve Jobs presentation prompt hook is derivative and the final recommendation to stick with specialised tools is entirely predictable. No fresh mental models or counterintuitive arguments appear.
I asked it for a 16 page PowerPoint presentation. Now the beautiful thing about this is this is a prompt that I use to have Steve Jobs effectively help me create an Apple style presentation
I would still use tools that are specialized like Gamma, which is agentic in nature as well, deep research
There is no guest; this is a solo monologue. The host presents himself as an AI transformation specialist but demonstrates only surface-level practitioner knowledge - tool-testing a consumer product release - rather than scaled operational experience or unique domain depth.
my name is Ryan Staley and I specialize in AI transformation specifically for go to market teams, executives. You name it, I'm nailing it
I'm a big proponent of like, all right, what else can I try this
The Vista Equity portfolio-scraping use-case provides genuinely concrete field detail - specific data fields requested, credit consumption numbers, and PE investment-horizon context - but much of the rest of the episode relies on vague qualitative impressions ('pretty solid,' 'some slides were good and some mediocre') with no reproducible metrics.
I asked for tabular format and... a portfolio company called Vista Equity... Brief description, Headquarters location, status of investment, Date of investment LinkedIn URL Number of salespeople, Chief Revenue officer and the CROs datehire
I burned through maybe 10,000 credits, which I think you get 20,000
This is an unstructured solo screencast monologue with no interviewing, no follow-up questions, and no analytical challenge to any claim. The host frequently loses his train of thought mid-sentence and the narrative is disorganised throughout.
I said basically twice. It's identifying like, like the revenue loss annually
I don't know what's going on there. So let me stop sharing temporarily and we'll get that back on in a second
Computed from the transcript - who did the talking, and the words that came up most.
Your competitors are already using AI. Don't get left behind. Weekly strategies used by PE Backed and Publicly Traded Companies → - In this video, Ryan Staley explores Genspark, a new AI super agent, by conducting three tests to evaluate its capabilities in presentation creation and research. He shares insights on the effectiveness of Genspark compared to other tools, discusses best practices for maximizing its use, and concludes with thoughts on its current limitations and potential for future development. Chapters 00:00 Introduction to Genspark: The New Super Agent 02:51 Testing Genspark: Presentation Creation 06:00 Research Capabilities: A Comparative Analysis 08:54 Maximizing Genspark: Best Practices and Strategies 11:45 Conclusion: Genspark's Current Limitations and Future Potential
Transcribed and scored by The B2B Podcast Index.
Speaker A: Genspark is being touted as the next new super agent, and my goal is to find out today with three tests if it can replace your entire tech stack. Those of you who don't know me, my name is Ryan Staley and I specialize in AI transformation specifically for go to market teams, executives. You name it, I'm nailing it. But today I want to go through this new super agent that's been released called genspark. And there's been getting a lot of noise, a lot of positive feedback, and what I'd like to do, because there's so much hype and money spent on branding, is always test out the tools and find out is it legit or should you quit using it? Right? And so I'm going to break down three simple tests that I did today, and by the end of the video, if you stay, I'll show you how you could use any agent and give you the prompt as well to execute and really, truly understand how to get the most out of any super agent. Okay, so let's begin. So today, let me share my screen. I'm going to get right into it and what you're going to see is the natural screen and flow that this has, which is really interesting. I don't quite understand why they have like this news feed below. I don't know what this is about. It's kind of interesting but also distracting at the same time. So. But what it's talking about are the examples of, as you can see, AI slides, AI sheets, you generate videos, call for AI chat, and all agents. Okay? So if you look at it, these are the areas along the side. Now what I'm going to do is like, I look at utility, right? So if I'm in business and I want to leverage this, being a entrepreneur, a CEO or for my clients, right? Sales leaders, marketing leaders, sales executive marketers, you name it. Like, I want to make sure this actually works for what I'm going to use it for, not just some fancy demo. And so what I did is I put it through three tests and I'm going to go through each one of these today. Okay, so let's start off with the presentation. All right, uh, so one of the things I did is, as you can see here, let me go all the way to the top. I created a prompt and what I did is I asked it for a 16 page PowerPoint presentation. Now the beautiful thing about this is this is a prompt that I use to have Steve Jobs effectively help me create an Apple style presentation, right? What he used to do and just mesmerize people. There's a whole book called the Presentation Secrets of Steve Jobs. Harvested some details from that integrated in this prompt. And here's what I'm asking too. And feel free to drop it in the comments if you want. Like I'm strongly considering any kind of prompts that I have within here to include those as a value for you. So if you want I could always include this. But drop it in the comments and let me know. So anyways, as you could see, this is very involved. It goes through all the visuals, uh, how to create this, how to psychologically engage. And so it's a pretty detailed output. And so what it does, I put it through and it took about 5, 6 minutes to create this entire deck. Let me show you what it came up with. And by the way, after it went through, looked like there were some slides that were good and some that were mediocre. So I asked it to redo it. So what this is is based off my value prop and so there's some unique things like work less, sell more. Okay, so that was strong. Uh, the villain, the revenue drag. And I kind of like how it's like dropping the calendar into the time glass. Um, pretty creative in terms of the animation there now, cost of inaction. Um, I like the ticking clock however. So basically it's basically identifying. I said basically twice. It's identifying like, like the revenue loss annually based on just the time burn of not implementing and integrating AI in your go to market team. Okay, so this is cut off. A little disappointed on that. This visual is pretty cool. It's got the uh, like almost kind of like the, the readout with the heads up display and it's got the teammate. However, like as you can see it's off center. This one's downloading. I don't know quite what's going there. This 60 second blueprint that was working before had a countdown clock. Holy smokes. Rule of three, time, intelligence and capacity. Right? Cyborg, superhuman and agent. This, this, I mean there's some elements this I like. However, like as you can see the $1 turning to 6 isn't showing up. This, this slide is fire. I think this is, this is fantastic in terms of the concept. And uh, this is more like you could insert or paste it in there with like some piro slides. Okay, so like I think this is solid. Has some really promising ones. Like I love this clock. Like it really gets across the component. Some of these other areas are cool but like, so like I'M a big proponent of like, all right, what else can I try this? And they do a comparison. Right. So the tool that I use is Gamma. So I threw this in Gamma. Basically the same exact output. I didn't change anything and really didn't even give it an extensive details. I just used the prompt and so this is what it came up with. Right. And I'm going to walk you through it because I was pretty impressed with Gamma with the output that I created and of course now that I'm trying to show it to you, it's not coming up. I don't know what's going on there. So let me stop sharing temporarily and we'll get that back on in a second. As you can see. Oh, here we go. As soon as I stop. It's right there. Totally right there. Okay, so let's show this here. All right, so, so it's pulling up. So if you haven't used Gamma, it is really strong for visuals and can create presentations, websites, social files and for some reason it's okay. There we go. So now this kind of walks through, same prompt all uh, Right. Doesn't have as much animations but like I actually like this better. I think the quality of the output was significantly better as you go through it. Okay. Ah, so as you can see it's got all these different areas that it's hitting on. It's got like that cyborg, like that head up display. I thought that. I think this looks so cool like with a little kind of Iron man. It's stayed truer to message that I was trying to get talks about the levers. I like this evolution journey from like cyborg to secret agent. And so in terms of round one, I would say like I would go with this to Gamma. I wouldn't use genspark for their presentations. I tried it one other time, really good at creating visuals but the visuals didn't export properly into PowerPoint. So. Okay, so let's look at number two. So number two is what comes up is research all the time. And you know there's a lot of research tools. So I think it's getting harder and harder for general purpose agents to compete with research tools like deep research across Google or across ChatGPT. And so however I saw their data release and I was really excited about it. So I'm like I have to try this. And the data release showed visualization. It's showed pulling up data was really, really impressive. Uh, so what I did is I gave it a very specific prompt and I didn't want to leave it like a lot of latitude by design so that it would give me the exact details and output I looked at. So if you look at this here, I asked for tabular format and, and I'm looking at a portfolio company called Vista Equity. If you don't know Vista Equity, they're one of the largest private equity companies in the world. I clicked on their portfolio company page and you'll see this with private equity or venture capital. And I asked them to click all the portfolio companies. Brief description, Headquarters location, status of investment, Date of investment LinkedIn URL Number of salespeople, Chief Revenue officer and the CROs datehire so I went through this and like this thing burned up a ton of my credits because it started to do a good job. However, it only pulled up 10. And what it got stuck on was the number of salespeople. So if you're a genspark provider or I should say you work at the company, you might want to look at like some fail safe if it can't find the details because it just kept looping and looping and looping and I had to stop it. Okay. At the same time I looked at like my credits and I had to upgrade. I burned through maybe 10,000 credits, which I think you get 20,000, um, just from these two exercises or three exercises I did. And I think this was the one that burned the most credit. So I was hyper specific and I'm like, all right, what happens if I try this in Gemini, right? So let's see if this. And so the one that I did in Gemini is Gemini has unlimited deep research. So I played put the exact same prompt in and as you can see, it pulled not just 10, but every single company in their portfolio with their, their LinkedIn URL. For the most of them, I should say about half of them. But then if you look over here. Let's go up. I'm sorry, let's go up. See this? Okay, so if we go over, it's got the whole name of the company, the website, the URL and the uh, company LinkedIn website. So real interesting. It also identified the acquisition date or the active date because when you're basically targeting like a portfolio company, the date of the investment's important because that basically fuels how active they need to be to get to an end result before they sell. That's usually three to five years or five to seven. Just a little extra. I'm giving you just based on my experience and what I've seen. So step two, I was a little Disappointed on that, uh, because I was looking for more. So basically Deep Research beats it on the Deep Research, if you will. That's the name. So let's shuffle on over. I want to go to, um, this is the prompt that I said at the end I wanted to provide for you and I could provide all three of these. The Steve Jobs prompt, the data analysis prompt, like the hyper specific research. Um, and at the same time, this last one I'm going to show you but comment in the chat thread and uh, if you, you know, if there's enough interest, then I'll, I'll create a page and we'll, we'll start to get this to you out on a reoccurring basis. Okay, so here's another one. Um, what are the top 10 ways you can use as a super agent? I'm the CIO and I basically list. I'm an AI transformation company who serves these. Um, I work with the CRO or CMO. My goal is free up 20 hours a week while also 2xing revenue in the next 12 months. So what I'm doing here, and this is critical for any AI tool to use and what you could do is basically ask the AI tool, how do I use you best? What am I going to get the best results with? And if you do that, uh, you'll start to get really good outputs because it'll, it'll process and think through and share with you how to do it. Now you can't ask Excel how to use Excel better. You can't do the same thing with PowerPoint either. Right? So what you're going to see on here is I have this where it gave me some pretty solid ideas like content creation, creation, distribution research, roadmap generator, market intelligence dashboard. I didn't get a chance to stay with this. This probably would have been pretty, pretty solid to do sales enablement pipeline, workshop data analysis, and even thought leadership amplification. So some really strong categories there. Now I said, all right, let's go deeper on one. So then the thing that I loved is it started to use all different tools and it started to look in areas for content creation and pick ideas. Right? So I think this was pretty strong in terms of what it was starting to do. And what it did is it then created a content strategy for exactly what I'm doing. It talked about my content pillars, which you see here, content mix, posting schedule. So, so I think like these are really strong foundational elements. And then it even started to create posts, uh, like a carousel post. So I'm going to run these through, put these in text and see what they look like when I visually generate them. And some of this looks like pretty solid language. Now it's obviously not to the level where I would want it for something that I would use. It's got a data visualization post, case studies, interactive. And so I would say the quality is pretty solid. And that goes through week by week, what I should do and then an engagement strategy and how I can measure success. Okay. Ah, so to summarize it all in all, I would say, you know, genspark, I have, I'm excited about the opportunity, but I don't think it's quite there yet in terms of the execution. I even tested with the exact prompt in the demo and it with uh, a like one minor tweak to customize it to me and it didn't come up with the right results and it got caught in the looping again. The reason why I'm sharing this video with you is there's a lot of marketing and a lot of hype about the perfect demo and shows really well. Like I said, I think genspark can be amazing potentially but right now it's not there for me. So I would still use tools that are specialized like Gamma, which is agentic in nature as well, deep research and um, that can be cross Gemini or Google. And then for ideation, it's okay for ideation but like right now I wouldn't necessarily latch myself to that wagon. I would still use a general purpose, large language models and then roll from there. So if you want to hear more from me, check out the next video. At the same time, uh, I will also include a link if you comment below about where you can start to get some of these prompts that I'm sharing because I think that's something that'll really start to maximize these videos for you, what you're doing and help you take things to the next level. So appreciate you joining me today and we'll see you all on the next video.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.